The psychometric properties of the Motivation Assessment Scale (MAS) were examined. Specifically, reliability, internal consistency, and construct validity were estimated using 90 ratings of different problem behaviors among 86 individuals with mental retardation. Data were analyzed under conditions of (a) three topographies of problem behavior and (b) two methods of calculation. Although reliability and internal consistency were generally poor, the results depended upon the above conditions. Factor analysis revealed that the four factors of the MAS differed from the subscale structure obtained in this study, proposed by the authors of the MAS. Thus, the construct validity of this scale was ambiguous. It is concluded that the MAS should be employed in conjunction with other instruments to assess the functional characteristics of problem behavior with individuals who have mental retardation.